Srabani Mukhopadhyaya

Localization control to locate mobile sensors

Localization is an important issue for Wireless Sensor Networks in a wireless sensor network. A mobile sensor may change its position rapidly and thus require localization calls frequently. It is important to control the number of localization calls, as it is rather expensive. The existing schemes for reducing the frequency of localization calls for mobile sensors uses the technique of extrapolation which involves simple arithmetic calculations. We propose a technique to control the localization that gives much better result. The proposed method involves very low arithmetic computation overheads. We find analytical expressions for the estimated error if the rate of localizations is specified. Simulation studies are carried out to compare the performances of the proposed method with the methods proposed by Tilak et al.

Multi-robot Assembling Along a Boundary of a Given Region in Presence of Opaque Line Obstacles

This paper presents a distributed algorithm for assembling a swarm of autonomous mobile robots on a common boundary of a given polygonal region in presence of opaque horizontal line obstacles. Robots and obstacles are initially scattered in an unknown environment and they do not have direct communication among themselves. The algorithm guarantees successful assembling of all the robots on the left boundary of the given region within finite amount of time and without facing any collision during their movement. The intermediate distances among the assembled robots are not fixed. In this proposed algorithm, the robots follow the basic Wait-Observe-Compute-Move model together with the Full-Compass and Synchronous/Semi-synchronous timing models.

Painting an Area by Swarm of Mobile Robots with Limited Visibility

This paper presents a distributed algorithm for painting a priori known obstacle free rectangular region by swarm of mobile robots with limited visibility capability. We have assumed that initially the visibility graph is connected. Our approach is to divide the region into some non overlapping strips, and to let each robot to paint one of these strips assigned to it. Width of the strips may vary for different robots. In the proposed algorithm, the robots follow the basic Wait-Observe-Compute-Move model together with the Asynchronous timing model.

Optimal Tree Search by a Swarm of Mobile Robots

Swarm robotic research in discrete domain assumes that robots in the swarm are randomly deployed over any graph and the robots can move only through the edges of the graph. In target searching, all the robots in the swarm are required to gather at the specially designated node, termed as target node. Moreover, during target search if the graph is guaranteed to be explored completely, the scope of the solution increases. An algorithm for tree searching by swarm of asynchronous robots of limited visibility has been proposed in this paper. An O(1) memory is assumed to be attached to each node of the tree. The target node is initially visible to at least one robot in the swarm. However, if it is executed on synchronous system, the algorithm takes O(n) computational cycles to gather all the robots at the target node after exploration of the tree completely, where n is the number of nodes in the graph.

Gathering in the Discrete Domain: State of the Art

Discrete domain swarm robotics is an emerging and challenging field of research. Unlike continuous domain where working place of the robots is a two-dimensional plane, in the discrete domain their working place is modelled by a graph. Robots are deployed on the nodes of a given graph, and they are allowed to move only along the edges of that graph. Consequently, the models used in the continuous domain are not always applicable in the discrete domain. Exemplified by the gathering problem, this article critically reviews models, assumptions, and approaches that have been proposed in solving the problems in the discrete domain.

Suitability of CORDA Model for Area Coverage Problem

COopeRative Distributed Asynchronous model or CORDA model is a basic computational model in the field of robot swarm. The objective of this work is to justify popularity and suitability of the CORDA model as a basic model of computation vis-a-vis other computational models. The problem of covering a target area has been taken up as an area of focus. This chapter presents a critical review of the various solutions of the coverage problem under CORDA model.

Classification of Robot Swarm Models in Discrete Domain

An emerging and challenging area of research in swarm robotics is to consider swarms deployed in discrete domains. In the continuous domain, it has already been established that different computational and behavioral models of the robot swarm play an important role in solvability of different fundamental problems. Due to some basic differences, not all existing models in the continuous domain are relevant or significant with respect to the discrete case. In this paper we draw an analogy between the models already existing in both the domains and propose a few relevant models for the discrete domain.

Cache oblivious algorithm of average filtering in image processing

Cache oblivious algorithms are cache conscious and cache efficient algorithms, independent of cache size of the computer system. We consider the simple average filtering algorithm in the field of image processing. This algorithm is used for image enhancement technique. The algorithm incurs huge number of cache misses while it processes the image matrix. We propose a cache oblivious version of average filtering algorithm. The new algorithm makes better utilization of cache than does original average filtering algorithm. The experiments indicate that the new algorithm performs better reducing the cache misses by at most 80% than its traditional counterpart being independent of cache size for various input matrices.

Genetic algorithm for double digest problem

The strongly NP-complete Double Digest Problem (DDP) for physical mapping of DNA, is now used for efficient genotyping. An instance of DDP has multiple distinct solutions. Existing methods produce a single solution, and are slow for large instances. We employ a type of equivalence among the distinct solutions to obtain almost all of them. Our method comprises of first finding a solution from each equivalence class by an elitist genetic algorithm, and then generating entire classes. Notable efficiency was achieved due to significant reduction in search space.

A 2-D Random Walk Based Mobility Model for Location Tracking

Performance analysis of different location tracking schemes needs a model that reflects the mobility of the mobile terminals in a realistic way. We propose a two-dimensional random walk based model with inertia for performance analysis of different location management schemes with specific update and paging mechanism. We used our mobility model to analyze the cost of movement-based location update technique with selective paging. We extend the model to have different values for inertia of motion and inertia of rest.